The purpose of early phase CAM research is to determine whether a particular CAM treatment shows sufficient indication of efficacy for further research on that treatment. Phase III clinical trial criteria are often applied in these situations, despite the fact that the purpose of a Phase III trial is to establish effectiveness as the basis for treatment recommendations. Since Phase III criteria are generally stringent, the expected result is that research on promising CAM approaches will be terminated prematurely. It is possible that shifting the criteria for early phase research could reverse this trend, with the ultimate result that a larger number of CAM therapies would be given appropriate amounts of later-phase research attention, so that the general population would eventually have greater access to research-proved beneficial CAM therapies. This application proposes methodological research on four approaches to improving early phase CAM research designs. (1) Randomization to treatment is the standard in Phase III studies, but in smaller early phase studies it results in considerable lack of balance between treatment and control groups, with subsequent bias and/or loss of statistical power. Design-adaptive randomization guarantees balance, but the statistical and clinical trial communities have taken the attitude that there it has analysis problems. The first aim is to test this claim; a showing of no analytic problems would then allow balanced randomization in small trials. (2) The intent-to-treat paradigm forces researchers to pay a price (in sample size and conservative bias) for having dropouts. The second aim is to establish the statistical characteristics of an alternative strategy for dealing with dropouts, which could greatly diminish bias, and might improve power. (3) An alternative version of hypothesis testing, call separation tests, has been proposed by the PI for reducing the excess type II errors resulting from using Phase III statistical criteria in Phase I studies. The third aim proposes an evidence-based review of CAM literature to estimate the impact that separation tests could have had on the developing programs of CAM research in multiple areas, with the intent of strengthening the argument for the use of this new kind of testing in early phase studies. (4) Whole systems research (WSR) is difficult in part due to the degree to which the practitioners can practice with fidelity to their system. A modification of the RCT appropriate for WSR has been proposed by the PI, and in this Aim designs using this modification will be created and tested. The overall purpose of this project is to strengthen early-phase CAM research and WSR across a broad spectrum of modalities, by justifying and popularizing methodological improvements. Since most CAM and WSR research takes place in early phase studies, the impact on appropriate testing of CAM could be considerable.